variation and used the corrected expression values in the software DESeq2 [52] to determine differential gene expression. The entire gene list with corrected expression values can be found on Additional file 6. After filtering out low expressed genes (mean FPKM < 1 across 19 samples), we identified 782 differentially expressed genes: 503 increased in the SZ samples and 279 decreased, at nominal p < 0.05 (Additional files 7 and 8, respectively) (Fig. 1a). Because of the relatively moderate sample size and experimental variation, only a small number of differentially expressed genes were statistically significant after correcting for multiple testing (42 genes by FDR < 0.05). Nevertheless, based on the FPKM values and expression changes for genes in the 22q11.2 deleted region, we considered p < 0.05 a reasonable threshold for calling differential expression. As seen in Fig. 1b, 36 of the 47 protein-coding genes in the 22q11.2 region, including most of the candidate genes implicated in the psychiatric manifestations or endophenotypes associated with 22q11.2DS (DGCR8, DGCR2, RANBP1, RTN4R, and COMT, for example) showed a significant decrease (~2-fold reduction, FDR < 0.05) in the patient samples compared with controls. One exception is CLDN5, which also showed larger than 2-fold difference in
